AdaGrad-Diff: A New Version of the Adaptive Gradient Algorithm
This work addresses the problem of stepsize tuning in optimization for machine learning practitioners, but it is incremental as it builds on the well-established AdaGrad framework.
The paper tackles the sensitivity of gradient methods to stepsize tuning by proposing AdaGrad-Diff, an adaptive algorithm that uses cumulative squared norms of gradient differences instead of gradient norms, resulting in improved robustness over AdaGrad in several practical settings.
Vanilla gradient methods are often highly sensitive to the choice of stepsize, which typically requires manual tuning. Adaptive methods alleviate this issue and have therefore become widely used. Among them, AdaGrad has been particularly influential. In this paper, we propose an AdaGrad-style adaptive method in which the adaptation is driven by the cumulative squared norms of successive gradient differences rather than gradient norms themselves. The key idea is that when gradients vary little across iterations, the stepsize is not unnecessarily reduced, while significant gradient fluctuations, reflecting curvature or instability, lead to automatic stepsize damping. Numerical experiments demonstrate that the proposed method is more robust than AdaGrad in several practically relevant settings.